gpt4free/g4f/Provider/Ails.py

106 lines
3.6 KiB
Python
Raw Normal View History

from __future__ import annotations
2023-07-28 13:07:17 +03:00
import hashlib
import time
import uuid
import json
2023-07-28 13:07:17 +03:00
from datetime import datetime
from aiohttp import ClientSession
2023-07-28 13:07:17 +03:00
from ..typing import SHA256, AsyncGenerator
from .base_provider import AsyncGeneratorProvider
2023-07-28 13:07:17 +03:00
class Ails(AsyncGeneratorProvider):
2023-08-27 18:37:44 +03:00
url: str = "https://ai.ls"
working = True
2023-07-28 13:07:17 +03:00
supports_gpt_35_turbo = True
@staticmethod
async def create_async_generator(
2023-07-28 13:07:17 +03:00
model: str,
messages: list[dict[str, str]],
stream: bool,
proxy: str = None,
**kwargs
) -> AsyncGenerator:
2023-07-28 13:07:17 +03:00
headers = {
"authority": "api.caipacity.com",
"accept": "*/*",
"accept-language": "en,fr-FR;q=0.9,fr;q=0.8,es-ES;q=0.7,es;q=0.6,en-US;q=0.5,am;q=0.4,de;q=0.3",
"authorization": "Bearer free",
"client-id": str(uuid.uuid4()),
"client-v": "0.1.278",
2023-07-28 13:07:17 +03:00
"content-type": "application/json",
"origin": "https://ai.ls",
"referer": "https://ai.ls/",
"sec-ch-ua": '"Not.A/Brand";v="8", "Chromium";v="114", "Google Chrome";v="114"',
"sec-ch-ua-mobile": "?0",
"sec-ch-ua-platform": '"Windows"',
"sec-fetch-dest": "empty",
"sec-fetch-mode": "cors",
"sec-fetch-site": "cross-site",
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/114.0.0.0 Safari/537.36",
"from-url": "https://ai.ls/?chat=1"
2023-07-28 13:07:17 +03:00
}
async with ClientSession(
headers=headers
) as session:
timestamp = _format_timestamp(int(time.time() * 1000))
json_data = {
2023-07-28 13:07:17 +03:00
"model": "gpt-3.5-turbo",
"temperature": kwargs.get("temperature", 0.6),
"stream": True,
"messages": messages,
"d": datetime.now().strftime("%Y-%m-%d"),
"t": timestamp,
"s": _hash({"t": timestamp, "m": messages[-1]["content"]}),
2023-07-28 13:07:17 +03:00
}
async with session.post(
"https://api.caipacity.com/v1/chat/completions",
proxy=proxy,
json=json_data
) as response:
response.raise_for_status()
start = "data: "
async for line in response.content:
line = line.decode('utf-8')
if line.startswith(start) and line != "data: [DONE]":
line = line[len(start):-1]
line = json.loads(line)
token = line["choices"][0]["delta"].get("content")
if token:
if "ai.ls" in token or "ai.ci" in token:
raise Exception("Response Error: " + token)
yield token
2023-07-28 13:07:17 +03:00
@classmethod
@property
def params(cls):
params = [
("model", "str"),
("messages", "list[dict[str, str]]"),
("stream", "bool"),
("temperature", "float"),
]
param = ", ".join([": ".join(p) for p in params])
2023-09-18 00:23:54 +03:00
return f"g4f.provider.{cls.__name__} supports: ({param})"
2023-07-28 13:07:17 +03:00
def _hash(json_data: dict[str, str]) -> SHA256:
base_string: str = "%s:%s:%s:%s" % (
json_data["t"],
json_data["m"],
"WI,2rU#_r:r~aF4aJ36[.Z(/8Rv93Rf",
len(json_data["m"]),
)
return SHA256(hashlib.sha256(base_string.encode()).hexdigest())
def _format_timestamp(timestamp: int) -> str:
e = timestamp
n = e % 10
r = n + 1 if n % 2 == 0 else n
return str(e - n + r)